Optimove Insights published its Holiday Shopping Report 2026 on 5 August 2026, reporting that 72 percent of the consumers it surveyed regularly or occasionally use AI tools such as ChatGPT, Claude or Gemini for shopping ideas and gift recommendations. The same report found that 53 percent of those consumers intend to shop only at retailers they used last year, which is the number that complicates the headline.
Two things need stating before any of the figures are quoted onward. This is vendor-published research, produced by a marketing software company about a market it sells into, not independent academic or trade research. And the sample is 648 United States consumers aged 18 and over with household incomes of 75,000 dollars or more, surveyed in Spring 2026. That is a US, higher-income sample, and it is not a general population.
What Optimove reported on 5 August 2026
The report, distributed through GlobeNewswire on 5 August 2026, covers how a group of US shoppers say they expect to approach the 2026 holiday season. Optimove disclosed its methodology, which is more than many vendor surveys do: 648 respondents, all US, all aged 18 or over, all in households earning at least 75,000 dollars a year, with fieldwork carried out in Spring 2026.
The framing figure is the 72 percent who said they use AI assistants for shopping ideas and gift recommendations either regularly or occasionally. Optimove chief executive Pini Yakuel is quoted in the release saying: "AI is no longer a future layer of holiday shopping. It is already part of how consumers discover products."
Where AI actually appears in the shopping trip
The report breaks the 72 percent down by feature, and the split is more useful than the headline because it shows which AI functions people say they reach for. The lead-in matters here: these are self-reported usage rates from the 648-person sample, not measured events in anyone's analytics.
| AI feature | Share of surveyed consumers using it |
|---|---|
| Product recommendations | 55 percent |
| Chatbots | 41 percent |
| Visual search | 31 percent |
| Personalised deals | 26 percent |
| Size and fit recommendations | 23 percent |
The ordering is worth reading closely. Recommendation is the dominant use at 55 percent, and it is the least transactional of the five: someone asking what to buy has not yet decided where to buy it. Size and fit at 23 percent is the most transactional, and it is the least used. The report does not claim a causal link between these, and neither does this article.
The trust figure, and the barrier sitting behind it
Optimove reported on 5 August 2026 that three quarters of the surveyed consumers, 75 percent, said they trust AI-generated recommendations as much as or more than other sources. The comparison sources are not broken out in what the release covers, so it is not possible to say from this data whether AI is being trusted more than a friend, a review site, or a retailer's own product page.
Against that, 37 percent named data privacy as a barrier to wider adoption of AI assistants. Those two figures can be held by the same person without contradiction: trusting the quality of an answer is a different question from being comfortable with what the system knows about you.
The tension the report does not resolve
This section is analysis, clearly labelled as such. Optimove reported that 55 percent of the sample expect to spend more than last year and 30 percent expect to spend about the same. It also reported that 53 percent intend to shop only at retailers they used last year, with 47 percent open to exploring new brands.
Put those beside the 72 percent AI usage figure and the honest reading is narrower than the headline suggests. If most of this sample uses AI to generate ideas but a majority still plans to buy where they bought before, then AI visibility is doing most of its work inside relationships that already exist. It helps a shopper decide what to get from a retailer they were going to use anyway. Converting a genuinely new customer through an AI answer is the harder and rarer event, and nothing in the report measures how often it happens.
That reading has a practical edge for anyone planning ecommerce marketing work around AI discovery. Being cited in an AI answer is worth having. Treating it as a new customer acquisition channel with the economics of paid search is a claim this data does not support.
What decides the purchase once the shopper arrives
The report also asked what factors drive the purchase decision. These are stated preferences from the same 648-person sample, and stated preference on price is a notoriously soft measure, since almost nobody tells a survey that price matters more than quality.
| Purchase factor | Share naming it |
|---|---|
| Quality | 81 percent |
| Price | 70 percent |
| Brand reputation | 46 percent |
On retailer preference, Amazon led at 81 percent, followed by Walmart at 71 percent and Target at 58 percent. All three are US retailers, which is a reminder of who was surveyed. On timing, 66 percent said they were willing to shop between July and October, while 63 percent said they may still wait for Black Friday and Cyber Monday. Those two figures overlap rather than compete, since the same person can be open to buying early and still hold back for a specific discount event.
The limits of this finding
A vendor survey can still be useful. It just has to be read with its constraints attached, and this one has four that matter.
- Single vendor, single interest. Optimove sells marketing software to retailers. The finding that AI-driven discovery is now central to holiday shopping is a finding that supports the market its products serve. That does not make it false, but it is not neutral ground.
- Disclosed but narrow sample. 648 people is a workable survey size, and the disclosure of geography, age floor and income floor is a point in the report's favour. The 75,000 dollar household income floor removes lower-income US shoppers from the picture entirely.
- Stated intent, not measured behaviour. Every number here is what someone said they do or plan to do. Purchase intent surveys and actual Q4 sales data routinely disagree.
- Timing gap. The fieldwork was done in Spring 2026 and asks about a Q4 that has not happened yet. Several months of price movement, product releases and AI assistant changes sit between the answers and the season they describe.
What the source does not say
The release does not report which AI assistant the respondents used most, or whether usage differed by assistant. It gives no breakdown by age, gender or region within the United States. It does not say what share of AI-assisted shopping journeys ended in a purchase, or how those journeys performed against non-AI journeys. It does not compare the 2026 figures to an equivalent 2025 survey, so no trend can be read from it here. And it says nothing at all about Thailand, Southeast Asia, or any market outside the United States.
What this means for Thai marketers
Start with the boundary: the sample is entirely US, entirely aged 18 and over, and entirely from households above 75,000 dollars a year. Nothing in the report was measured in Thailand. Any Thai retailer quoting the 72 percent figure as a Thai statistic would be misrepresenting it.
What can reasonably be taken from it, as reasoning rather than as reporting, is the direction of the question a Thai retailer should be asking about their own market rather than borrowing an answer from this one. If a customer asks an AI assistant for a gift idea in Thai, does your product information exist in a form the assistant can read and repeat correctly, including price, availability and the Thai product name. That is the practical core of generative engine optimisation work, and it is answerable with your own site rather than with a US survey.
The 53 percent loyalty figure points the same way. If AI mostly assists people who already know a retailer, then the material that describes your products and answers real customer questions is doing double duty: it serves the existing customer and it is the raw material an assistant would quote. That is an argument for content marketing that answers questions precisely, not for volume.
FAQ
Does the 72 percent figure apply to Thai shoppers?
No, and the source does not claim it does. The survey covered 648 US consumers with household incomes of 75,000 dollars or more, so the figure describes that group and nobody else. There is no equivalent Thai figure in this report.
Is this independent research?
No. The report was produced and published by Optimove, a marketing software vendor, and distributed via GlobeNewswire on 5 August 2026. The methodology is disclosed, which helps, but the research was not commissioned or reviewed independently as far as the release states.
Do I need to change anything before Q4?
Nothing in this report requires an action, because it measures stated intent in a US higher-income sample rather than outcomes in your market. The reasonable move is to check whether your own product information is accurate and readable, which is worth doing regardless of what this survey found.
If most shoppers stay with familiar retailers, is AI visibility worth pursuing?
It is worth pursuing, but the expected return is different from what the headline implies. On this data AI discovery looks like it mostly assists existing relationships, so the realistic gain is defending customers you already have rather than a flow of new ones. The report does not measure new customer acquisition through AI at all.
Why do 55 percent expect to spend more while 53 percent stay with the same retailers?
Those two answers are not in conflict, because spending more and spending it in the same places are separate decisions. What the pairing suggests, as reasoning rather than as a source claim, is that a larger budget in this sample is more likely to be captured by incumbent retailers than redistributed to new ones.
The short version
Optimove Insights reported on 5 August 2026 that 72 percent of 648 surveyed US consumers in higher-income households use AI tools for shopping ideas, while 53 percent plan to buy only where they bought last year. It is vendor research on stated intent in a single market. The release is on GlobeNewswire.
If you want to work out what AI discovery actually looks like for your own products in the Thai market, rather than borrowing a US number, our team is happy to take a look with you.
